"""OPDS server backed by FreshRSS. An OPDS 1.2 catalog whose shelves mirror FreshRSS categories and feeds, serving one EPUB per article, generated in memory at download time. Nothing is written to disk: the catalog is rebuilt from the FreshRSS API as it is browsed, and a completed EPUB download marks the corresponding article read upstream. Layout, in dependency order: ``config`` The JSON document, its interpolation (``${ENV}``, ``${file:...}``) and its validation. Everything else is configured from here. ``urls`` Identifier and filename helpers: base64url for path identifiers, stable ``urn:uuid`` values for Atom ids, and safe download filenames. ``cache`` A bounded in-memory TTL cache with per-key single-flight, so a reader browsing quickly does not turn into a burst of upstream requests. ``xhtml`` Feed HTML into well-formed XHTML. Everything the reader's XML parser sees has been rebuilt by this module. ``images`` Media-type sniffing and dimension parsing, so a book's manifest is honest about its bytes and a cover can be chosen without an imaging library. ``fetch`` The one place that fetches a host the operator did not choose. Article images are fetched through an SSRF guard that dials a validated address. ``epub`` The EPUB 3 writer, assembled entirely in memory. ``freshrss`` The FreshRSS client, over its Google Reader compatible API. ``auth`` HTTP Basic authentication, with a throttle on failed attempts. ``opds`` OPDS 1.2 Atom document builders. ``server`` Routing, the catalog logic, and EPUB delivery. Run it with ``python -m app --config config.json``. """ __version__ = "1.0.0"